Skip to content

Market-Power Screen

Diagnostic screen — instantiates Bottleneck Power Governance

Tests whether an access point is genuinely a non-substitutable bottleneck — and pins down who controls exactly what — before any access duty is imposed.

A Market-Power Screen is the gating test that decides whether there is a bottleneck worth governing at all — and, if so, where. Its defining move is to separate a merely dominant or preferred provider from a genuinely non-substitutable chokepoint, and to refuse to trigger any access duty until that separation is made on evidence. It produces three findings and nothing else: whether the access point is truly hard to substitute, where control actually sits (which may not track formal ownership), and who depends on it and how captive they are. It is the front of the pipeline — a diagnostic, never a remedy.

Example

An internal-market team at a competition authority is asked whether a hyperscale cloud's proprietary managed-database service is a bottleneck that warrants access rules. The screen runs the substitutability question first: could customers realistically move to a rival within a reasonable time and cost? It tallies the frictions — proprietary APIs, data-egress fees, re-architecture effort — and applies a hypothetical-monopolist logic: if the provider quietly raised effective prices ≈10%, would enough customers leave to make it unprofitable? If not, the point is non-substitutable. Then it maps the control locus: the provider controls the API surface, egress pricing, and deprecation schedule, but not the underlying open query language — so the power boundary is the proprietary layer, not the standard beneath it. Finally it maps dependents: which downstream products would be stranded by a cutoff.

The output is a scoped finding: "non-substitutable across roughly a 2–3 year switching lead time; control sits at the proprietary API and egress layer; ≈40 downstream products dependent." That is what tells the governance regime whether to act and, crucially, on which layer. Had the screen instead found a close substitute a quarter away, its recommendation would be to impose nothing.

How it works

  • Test substitutability, not popularity. Apply hypothetical-monopolist reasoning to switching cost and lead time — a large market share with an easy exit is not a bottleneck.
  • Locate the control locus and its boundary. Identify what the controller can actually change — ranking, API, price, certification — separately from what it formally owns.
  • Map dependents and their exit options. Enumerate who rides on the access point and how captive each is; captivity, not headcount, is the signal.
  • Grade the finding. Attach an uncertainty band and a re-screen trigger, because substitutes and lock-in both move.

Tuning parameters

  • Substitutability horizon — how long and costly a switch may take before a substitute "counts." A short horizon finds more bottlenecks; a long one excuses more.
  • Power-boundary granularity — the whole firm versus the specific controllable layer; too coarse over-regulates, too fine misses bundled leverage.
  • Dependence threshold — how much of an actor's operation must ride on the point to count as captive.
  • Evidence bar — desk analysis versus switching surveys and market testing; a higher bar is slower but far harder to game.
  • Re-screen cadence — one-shot versus periodic, since a substitute can appear or a lock-in deepen.

When it helps, and when it misleads

Its strength is that it stops the whole regime from firing on a merely popular provider, and it aims every later duty at the exact controllable layer rather than the whole firm. Its difficulty is that substitutability is genuinely hard to measure, and the screen is gamed from both sides: a controller points to a distant, theoretical substitute; a complainant insists switching is impossible. Its classic misuse is being run backwards — assembling the market definition to reach a verdict already chosen, "bottleneck" or "not," to license or block intervention. The discipline that guards against this is the hypothetical-monopolist test done honestly, with the uncertainty carried forward and the screen re-run as the market moves rather than frozen at the first finding.[1]

How it implements the components

The Market-Power Screen fills only the diagnostic components — the ones a test can produce:

  • non_substitutable_access_point — its central output: the evidenced finding of genuine non-substitutability, distinguishing a bottleneck from a strong competitor.
  • control_locus_and_power_boundary — it pins where control actually sits and what the controller can change, apart from formal ownership.
  • dependency_and_affected_party_map — it enumerates who depends on the access point and how captive they are.

It scopes the problem but imposes no duty: the access conditions come from Essential Facility Access Rule, the price and non-discrimination duties from Common Carriage Obligation, and the live enforcement channel from Abuse Complaint and Appeals Process.

Notes

The screen deliberately sets a high bar: most preferred providers are not bottlenecks. Keeping that bar high is what keeps access duties rare and targeted — an over-broad screen turns every strong competitor into a regulated one, which is its own harm. Every downstream mechanism here should be able to point back to a screen finding that justifies it.

References

[1] The SSNIP or "hypothetical monopolist" test asks whether a candidate would profitably sustain a small but significant, non-transitory increase in price without enough customers defecting — the standard antitrust method for market definition. It is referenced here to keep "bottleneck" a measured finding rather than a label, not as a claim about any specific market. ↩